Battery lifetime has become one of the top usability concerns of mobile systems. While many endeavors have been devoted to improving battery lifetime, they have fallen short in understanding how users interact with batteries. In response, we have conducted a systematic user study on battery use and recharge behavior, an important aspect of user-battery interaction, on both laptop computers and mobile phones. Based on this study, we present three important findings: 1) most recharges happen when the battery has substantial energy left, 2) a considerable portion of the recharges are driven by context (location and time), and those driven by battery levels usually occur when the battery level is high, and 3) there is great variation among users and systems. These findings indicate that there is substantial opportunity to enhance existing energy management policies, which solely focus on extending battery lifetime and often lead to excess battery energy upon recharge, by adapting the aggressiveness of the policy to match the usage and recharge patterns of the device. We have designed, deployed, and evaluated a user- and statistics-driven energy management system, Llama, to exploit the battery energy in a user-adaptive and user-friendly fashion to better serve the user. We also conducted a user study after the deployment that shows Llama effectively harvests excess battery energy for a better user experience (brighter display) or higher quality of service (more application data) without a noticeable change in battery lifetime.

See also the UMass Llama page.

Collaborators

• Nilanjan Banerjee - UMass Amherst
• Mark Corner - UMass Amherst
• Ahmad Rahmati - Rice University
• Lin Zhong - Rice University
• Maria Kazandjieva - Stanford University
• Denitsa Tilkidjieva - Rutgers University
• Serene Chong - University of San Francisco
• Nitin Ramamurthy - University of San Francisco

Publications

• N. Banerjee, A. Rahmati, M.D. Corner, S. Rollins, and L. Zhong, "Users and Batteries: Interactions and Adaptive Energy Management in Mobile Systems", in Proceedings of the 9th International Conference on Ubiquitous Computing (Ubicomp 07), Innsbruck, Austria, September, 2007.
• D. Tilkidjieva, N. Banerjee, M. Kazandjieva, S. Rollins, and M. Corner, "LLAMA: An Adaptive Strategy for Performing Background Tasks on Mobile Devices", in poster proceedings of the 7th IEEE Workshop on Mobile Computer Systems and Applications (WMCSA 2006), Semiahmoo Resort, WA, USA, April, 2006.

Analyzing and Predicting Mobile Device Usage Patterns:In recent years, there has been a significant focus on developing distributed systems to run on mobile computers such as laptops. However, experimental analysis of such systems often focuses on synthetic traces of mobility and usage patterns. Over the past several months, we have undertaken a study to gather data about real usage patterns of laptop computers. We deployed a measurement tool on over 60 laptops that periodically records information about the state of various resources including the battery and the network connection. We have several months of data that we have analyzed in order to classify the usage patterns of the mobile users in our study.

Our previous work used this data to make a case for adaptive energy management. Our analysis focused on battery usage and recharge patterns and our results indicated that users frequently charge their batteries with a high percentage of the charge remaining. Further, we concluded that recharge patterns differ greatly among users. The goal of this work is to extend our analysis to consider more general device usage beyond battery. We attempt to classify user behavior along several axes. Further, we consider whether we can accurately predict availability of devices and recharge patterns in order to better support mobile systems.

A Tool to Gather Data about Mobile Device Synchronization Patterns:Ensuring seamless anytime, anywhere access to data in an environment where personal mobile devices are becoming increasingly pervasive is a significant research and development challenge. Ideally, a user would be able to access any of her personal data on any of her personal devices at any time. To enable this functionality, the devices themselves must synchronize data in an automated way. However, the implementation of such an automated process is not straightforward. It must synchronize as often as possible to ensure consistency of data, but must also be careful not to exhaust the resources, such as energy, of the participating devices.

To guide the development of such an algorithm, we are developing a tool that can be used to gather information about the usage patterns of personal mobile devices. The goal of the tool is to identify when users use their devices, when they typically synchronize their devices, and how they typically drain the batteries on their devices. We can then use this information to determine the best methods for automatically synchronizing data.

Because most consumer devices, such as cameras and mp3 players, do not have yet a wireless interface, our tool works by gathering data when a device is connected to a PC for synchronization. For example, when a digital camera is connected to a PC to upload photos, our tool detects the USB device and retrieves as much information as possible about the device and its content. Our tool can also prompt the user for information by displaying a pop-up window that asks a series of questions about the device currently connected.

Collaborators

• Cheryl Chang-Yit - University of San Francisco
• Priyanka Daultani - University of San Francisco
• Nilanjan Banerjee - UMass Amherst
• Mark Corner - UMass Amherst

The goal of this project is to move toward better integration of and more seamless communication between personal mobile devices. Increasingly, a single user is responsible for managing several mobile devices. These devices may include one or more desktops, a laptop, a mobile phone, and any number of other consumer electronic devices such as cameras, mp3 players, DVRs, and game consoles. Additionally, it is quite common to access the same content on several devices. For example, many users transfer mp3s from their desktops to their iPods every morning. Similarly, users often transfer content from their desktops to their laptops to enable mobile access. This manual process is both cumbersome and relies on users to remember on which devices their content is stored since there is no good way of determining that information short of looking at the file systems of every device.

Our personal metadata repository is a web-based application that enables the user to visit one web page and see an integrated view of the content stored on all of her mobile devices. It is essentially a web-based view of a distributed file system. When a device is connected to the server, it uploads information about its current directory structure. It may also view information about the directory structures of all other devices. Building on Google Gears, we also enable offline access to the application so that a device may download the distributed file system view and access it even when disconnected from the network. This enables a user to easily determine where the latest versions of her content are stored.

Future extensions to this application include support for a broader range of devices including consumer electronics, support for device-to-device communication that does not require a server, management of the underlying data, and integration of energy-aware algorithms.

Collaborators

• Tushar Pednekar - University of San Francisco

As devices grow smaller and mobility becomes a reality, energy management is becoming a serious concern. Current approaches to manage energy, including dynamic voltage scaling and adaptively turning off unneeded components, only allow current laptops to remain active for up to 8 hours on a single charge. Other mobile devices, such as PDAs, are similarly limited in battery lifetime.

In order to conserve energy, devices are often put into a low-power suspension or hibernation state. While in these low-power states devices are unable to communicate and are unaware of available network resources.

The Hierarchical Power Management (HPM) project is a research effort aimed at allowing devices to remain aware of their environment with minimal impact on battery lifetime. Our approach is to integrate a hierarchy of independent platforms, optimized for different functionalities and workloads, into a single form factor. When not needed, the higher more power-hungry tiers can be powered down or suspended while lower tiers perform service discovery and small maintenance tasks. Our focus is to understand the practical realities of using hierarchical architectures and to provide OS-level mechanisms that effectively use such architectures to maximize battery lifetime.

Our two active projects are Turducken, a hierarchical approach to building mobile devices, and Triage, a HPM supporting software framework for tiered microservers.

See also the UMass Hierarchical Power Management page.

Collaborator

• Nilanjan Banerjee - UMass Amherst
• Mark Corner - UMass Amherst
• Deepak Ganesan - UMass Amherst
• Jacob Sorber - UMass Amherst

Publications

• N. Banerjee, J. Sorber, M. D. Corner, S. Rollins, and D. Ganesan, "Triage: Balancing Energy Consumption and Quality of Service in a Microserver", in Proceedings of The Fifth International ACM/USENIX Conference on Mobile Systems, Applications, and Services (MobiSys '07), Puerto Rico, June, 2007.
• J. Sorber, N. Banerjee, M. Corner, and S. Rollins, "Turducken: Hierarchical Power Management for Mobile Devices", in Proceedings of the Third International Conference on Mobile Systems, Applications, and Services (MobiSys 05), Seattle, WA, USA, June, 2005.